Chris Hall

Monash Centre for Synchrotron Science

Monash University, Melbourne, Australia.

Imaging small things Light microscopy

limited by penetrationLimited by wavelength (100s nm)

Electron microscopyHas to be done under vacuumExquisite resolution (0.2 nm)

X-ray microscopyDifficult to focus the x-rays15 nm is a very good resolution

Crystallography and coherent diffractive imaging (CXDI) X-ray crystallography

Details down to 0.1 nm…but only with crystals.

CXDI - Microscopy without lensesUse the coherence of the illuminating light to

obtain structural information.All the advantages of x-ray microscopy, but

much higher resolution (<10 nm)

The Goals of CXDI Imaging single (large) proteins Eventually use an x-ray free

electron laser Collect an image of the sample

(protein)…before it is blown apart.

More modest goals…small biological entities (cells).

70 nm

CXDI …how does it work?Axioms:

If the waves incident on the object are plane waves. Then in the far field, after interacting with the object the wave represents a Fourier Transform of the object.

To get back to an image of the object you simply need to perform an inverse Fourier transform.

In light optics the lens does this.

Detection and reconstruction If we can detect the complex far field wave

we can do the FT-1 with a computer. The problem is that our detectors only

detects intensity. No phase information is measured directly.

However, the phase can be estimated then refined by iteration.

Fourier transforms

Frequency

Intensity

F

Phase

Gershberg-Saxton (and family)

CXDI for real

J. Miao, K. O. Hodgson, T. Ishikawa, C. A. Larabell, M. A. LeGros and Y. Nishino, Proceedings of National Academy of Science of the USA 100, 110-112 (2003)

Fresnel CXDI

Fresnel CDI

Where will it happen

APS, Illinois, USA AS, Victoria, Australia

So, what are the current detector requirements?

In the geometry used now:

Energy range for > 50% DQE = 0.5 – 4 keVDynamic range in image (noise to saturation) >107

Pixel size ideally < 25 mPixel number >1024 x 1024Maximum local count rate 104 per pix per second

Trade studyCriterion Essential Desirable

1 Dynamic range 104 107

2 Pixelation 2048 x 2048 >4096 x >4096

3 Pixel Size <100 m < 50 m

4 Energy Range 0.5 keV to 3 keV 0.5 keV to 3 keV

5 Detective Quantum Efficiency

10% 50%

6 Rate Capability > 104 / pixel/ second > 106 / pixel/ second

7 Frame rate 1 Hz 1 kHz

8 Vacuum compatibility

10-4 bar 10-6 bar

9 Dead time fraction 10% 1%

10 Parallax error < 1 pixel at 20 degrees

11 Gaps in FOV Acceptable None

Commercially available detectors Easy option, but… Not designed for our purposes

XFEL pulse structure100 msecs 100 msecs

600 secs

Up to 30,000 pulses

100 fsecs104 photons, incident on the sample

99.4 msecs

Thank you for your attention

Thanks to my colleagues at MCSS and the CXSProf. Rob Lewis (MCSS)Prof. Keith Nugent (Melbourne University)Prof. Andrew Peele (La Trobe University)Dr. Garth Williams (Melbourne University)Dr. Mark Pfeiffer (La Trobe University)Evan Curwood (Monash University)Mr. Andrew Berry (Monash University)Dr. Wilfred Fullagar (Monash University)

The Australian Research Council fund theCentre of Excellence in Coherent X-ray Science.

Results so far

Gold blobs on Si3N4 substrate

Specifications for the CXS End Station apparatus Includes a 4 axis laser Doppler

displacement meter which provides:Positioning of sample w.r.t beam of 30 nmResolution in closed loop feedback of 0.2

nm=>Stability of sample w.r.t. beam of 2 nm